Stave Core Electrical Considerations Ned Spencer UCSC Sergio

Stave Core Electrical Considerations Ned Spencer (UCSC), Sergio Diez (DESY), David Lynn (BNL) Local Support Assembly, AUW Nov 3 2014 1

Electrical Isolators. 2. 275 mm OD Ti pipe internal to stave — 2. 5 mm OD pipe external Note both isolators are identical; there are no capillaries (0. 8 mm OD) into the stave Drawings by Peter Sutcliffe→ https: //twiki. cern. ch/twiki/bin/viewauth/Atlas/UKThermo. MECHDoc 2

Kapton Cable Proposal From Mike Dawson 1. Electrical insulates isolator’s nipples 2. NTC Two conductor cable Dual purpose Kapton cable proposal But only need to isolate this nipple 3

Motivation • Just before the September ITK meeting Mike wanted comments on his cable proposal as well as suggestions on how to ground the pipe to the foam • This led to discussions at the ITK about stave grounding and the electrical isolators • This talk is the beginning of the considerations of the stave core grounding scheme • Plan is for Ned, Dave, and maybe a few others to generate and short stave core grounding requirements document • There is some opposing considerations between wanting to “RF ground everything” (Ned) and keeping the stave core easy to build (Dave) • We expect these electrical grounding considerations will be very useful for the petals as well 4

Stave Lumped Element Model Preliminary model for discussion of necessary connections (e. g. transverse facing resistances not yet included) A B Boldface lines shows ground connections we think we need (via discussion with Ned) A. Ti pipe, foam, and facings to EOS. One location for each element only B. DC connection (e. g. pins) between adjacent foam pieces 5

Stave Lumped Element Model at 2. 5 MHz Note at 2. 5 MHz l = 120 m >> 1. 3 m stave length; lumped element model holds In blue are shown capacitive and inductive impedances at 2. 5 MHz, estimate peak of ABC 130 bandwidth. See backup slides for bandwidth estimate. 6

Inputs for Stave Lumped Element Model Resistance of K 9 Allcomp foam =. 15 W: Assumes rfoam ≈ 4. 3 x 10 -3 W cm • Measured at UCSC • Half length foam dimensions 15 cm x 5. 2 mm x 10 mm • Don’t yet know AC characteristics (Ned to study) Longitudinal resistance of facing = 0. 6 W: rface ≈ 4. 3 x 10 -3 W cm • Calculated for two layers of 45 gsm K 13 C 2 U fiber assuming electrical resistivity quoted in K 13 C 2 U data sheet of r. K 13 C 2 U ≈ 1. 9 x 10 -4 W cm. Should be measured. Resistor dimension 15 cm long , 10 cm wide, 40 mm thick. • Don’t yet know AC characterstics (Ned to study) Resistance of Ti Pipe = 0. 1 W: Assume r. Ti = 5 x 10 -5 W cm • Source: http: //asm. matweb. com/search/Specific. Material. asp? bassnum=MTU 020 • Resistor dimension, 15 cm long, 2. 275 mm OD, 100 mm wall • Ned measured 4. 3 x 10 -5 W cm Capacitance of foam to facing ~ 2. 5 n. F • 30 cm long sections, 50 mm hysol/BN thickness, 10 mm wide foam • dielectric constants of hysol and BN are 4 and 4. 5 respectively. Assume 4. 2 as average. 7

Things to Consider for Stave Lumped Element Model Capacitance of foam to pipe ~ 1 n. F • 30 cm long sections, 100 mm hysol/BN thickness, 10 mm wide foam • dielectric constants of hysol and BN are 4 and 4. 5 respectively. Assume 4. 2 as average. Inductance of 1 mm pin connecting foam pieces < 1 n. H • Complete guess for now. Should be small but need to consider what is the current loop. Model to be used to understand couplings and grounding 8

Pipe grounding • Our specially ordered pipe has a thinner than normal oxide on surface; it is electrically conductive • One option is to braze litz wire between nipple/pipe joint, but we haven’t found a small litz wire yet to try. Also, litz wire looses effectiveness > 3 MHz. • Maybe make a square metal collar (see figure) that allows a low inductance kapton/copper ground plane to be glue into seem. Collar is Ag-epoxy connected to pipe. Bonus is that NTC (thermistor) has a flat surface on which to mount. 9

Foam grounding • One possible idea • In any case grounding the foam should not be difficult 10

Simultaneous Foam Pipe Grounding + NTC for Interlocks Later make out of Titanium for CTE match with pipe? • Update. To avoid Ag-epoxy, Ned suggests maybe plating part of Ti pipe and then soldering a metal tab to it (replaces collar). • Is this possible? 11

Carbon fiber facing grounding • Resistance between layers of 0/90/0 layup now being investigated • Capacitance between layers should be measured as well • For now LBNL will investigate embedding of metal strip in EOS region during co-cure • Possible suggestion where metal runs through slits in pre-preg to contact all three layers 12

Foam to Foam Connection • Want to ground all foam pieces • So connect adjacent foam blocks with pins or some other mechanism • Ned suggests Kapton tubes with a sputtered nickel rather than pins to give larger surface contact area. Use epoxy to keep contacts airtight. Note: pipe not shown 13

Additional Connections? • Do we need to ground honeycomb? • Do we need to ground C-channels? • In both cases Ned suggests we should, but this may be difficult. We will study this • Also, do both the inlet and outlet side of pipe need grounding (rather than just one or the other? )

Next Steps • Continue to measure/characterize each element in terms of RF behavior • Better understand where/which DC connections are needed • Design best/simplest connection methods • Write up a short electrical requirements documents (Ned, Dave, others? ) • Determine if there is a good way to make useful measurements on an actual stave to verify expected RF behavior 15

Backup Slides 16

RF Frequencies of Interest • Assumption is most sensitive electrical element is the high gain input of the ABC readout chip • Have been unable to obtain frequency response of ABC chip from designers • So use a published ABC shaping function and make an LTspice model with a similar response function • Then generate a Bode plot from that LTspice model to obtain the frequency response 17

Circuit to Mimic ABC Impulse Response • 1 f. C impulse response shown • Used a ABC 130 paper to approximate the impulse response • Unlike ABC 130 this circuit inverts; easier to implement with standard LT Opamp models 1 f. C impulse response 18

Circuit to Mimic ABC Impulse Response : Frequency Response Bode plot 19

Circuit to Mimic ABC Impulse Response : Frequency Response on Linear Scale 100 k. Hz 30 MHz 2. 5 MHz 20

Related Note: Detector with Two Hybrid LTspice model Available if anyone wants it; email D. Lynn ABC Sub-circuit of previous slides Hybrid 21